The interest of this paper is the determination of the optical properties of oxygenated (saturation above 97 %) hemoglobin in clinical relevant concentrations (ranging from 5 to 15 g/dl), dependent on the layer thickness. Furthermore the generation of a high rate data set for training with machine learning approaches was intended. With a double integrating sphere setup (laser diodes from 780 to 1310 nm) - as a well referenced method - and flow through optical cuvettes ranging from 1 to 3 mm layer thickness, the transmission (MT) and reflection (MR) values of the samples were acquired. From those the layer thickness independent absorption (μa) and reduced scattering coefficients (μs') were calculated by the means of the Inverse Adding Doubling (IAD) algorithm. For each sample the same coefficients should result correspondingly for all cuvette thicknesses in test. This relationship serves as an internal standard in the evaluation of the collected data sets. In parallel a spectrophotometer in the range from 690 to 1000 nm recorded transmission spectra for all samples as a second reference. First, the IAD algorithm provided optical co-efficients (μa, μs') in all measurements, with few exceptions at low hemoglobin concentrations. The resulting coefficients match independently of the layer thickness. As a main second result, a high rate data set was generated which serves for further analysis - for example with machine learning approaches.